CAC payback
CAC payback is the point in a paying customer's lifecycle when the revenue they've contributed equals the cost of acquiring them. That cost is the total acquisition spend divided across paying customers, so each one's share covers a portion of the money spent reaching people who never converted. Before that point, the customer is a net loss for the business; after it, they're a net gain.
Definition
Industry-standard SaaS metric for how long it takes acquisition spend to be recouped through the acquired customers' revenue, typically computed as CAC ÷ ARPU.
The standard formulation
The standard formulation divides total customer acquisition cost by average revenue per user — sometimes with variants substituting ARPA, contribution margin per customer, or monthly gross margin per customer. A team reports "we recover our acquisition cost in N months."
The numerator aggregates acquisition spend across some period; the denominator aggregates revenue across the full paying base. Both blend across populations and time windows that don't correspond to each other, so the ratio carries units of months without measuring the payback duration of any specific cohort.
The aggregation-axis mismatch
Cohort revenue can be averaged two ways. One holds the cohort fixed and observes revenue accumulating over its own lifetime; this answers "how did this cohort pay back." The other holds the calendar moment fixed and sums across all cohorts present; this answers "what's the business's current revenue per customer." Standard CAC Payback uses the second axis as its denominator (aggregate ARPU) and presents the result as if it answered the first.
A team reporting "CAC payback is 14 months" cannot identify the cohort whose 14-month payback that figure describes. The number is computed from aggregate populations; no specific cohort's payback duration is being measured.
The aggregate ARPU denominator carries structural assumptions that don't apply to the cohort whose payback is supposedly being computed. New cohorts live in starting churn, where churn rates are highest and revenue retention is lowest; the aggregate ARPU pool is dominated by customers who survived that window. New cohorts also encounter current pricing and current plan distribution, while aggregate ARPU reflects the offering's cumulative pricing history and the upsell trajectory of long-tenured customers. The metric systematically misreads new-cohort payback because the denominator draws from a different population than the cohort being measured.
The CAC-offset trajectory
The positive formulation reframes CAC payback as a cohort-specific question: for this specific cohort, how long did cumulative revenue take to cross cumulative acquisition cost?
The operational form is the CAC-offset trajectory — cohort revenue plotted as a line accumulating against the cohort's own time, with the y-axis offset by the cohort's accumulated acquisition cost. The trajectory starts at a negative value (the spend has been made; no revenue yet) and climbs as revenue accrues. The moment it crosses zero is the cohort's payback duration, observable as a visual event in the cohort's revenue history.
Plotted alongside other cohorts' trajectories, the view foregrounds per-cohort divergence. Cohorts differ in slope (some accumulate revenue faster than others), in payback timing (when each crosses zero), and in destination (where each is heading by 5-year LTV). Each divergence is analytical signal — what changed between cohorts that produced the trajectory differences?
The two-stage correction
Stage one computes payback at conversion-cohort granularity, accumulating revenue for the cohort of customers who converted in a given month rather than relying on aggregate ARPU. The cohort's own paid revenue is what's being accumulated, which directly addresses survivor-skew and pricing-drift. Sophisticated SaaS finance practice often runs this stage, supported by tooling like ChartMogul, ProfitWell, or BI workspaces in Mode and Looker.
Stage two defines the cohort at the acquisition level rather than the conversion level — "users acquired by month X's spend, regardless of when they later converted." A user converting in October may have first encountered the offering months earlier, so October's conversion-cohort and October's acquisition-cohort aren't the same population. Acquisition-cohort framing directly answers "did this acquisition spend pay back" by aligning the measured population with the spend being measured. The apparatus required is attribution maintained across the signup-to-conversion lag, and stage two is rare even in sophisticated practice.
Example
A self-serve SaaS team reports CAC payback of 12 months using CAC ÷ ARPU. Computing it at conversion-cohort granularity reveals that the January 2024 cohort crossed CAC at month 19; the March 2024 cohort hasn't crossed at month 14, with the trajectory still climbing slowly; the June 2024 cohort is on track to cross at month 9 based on a steeper slope. The "12 months" was an artifact of averaging across cohorts at different lifecycle stages, with a denominator dominated by older survivors whose pricing and retention don't reflect what new acquisitions encounter. The divergence between cohorts is itself the operational signal — something compressed payback between January and June, and the trajectory surfaces it as a question the aggregate number couldn't ask.